Problem 2 (35 points): Nathan Armstrongs plan for the 50 smart valves had begun to crystallize

during his trip to the Blue Ridge Mountains. Upon returning, however, he was dismayed to discover
two new messages relating to the contract. The first was from Richard Hubble, the president of
Gemini Software and the next from Alicia Harrington, head of the Engineering Division. They had
several things to discuss. After talking to both, he realizes that the situation is quite different now.
There is no uncertainty in either the development of the control software or that valve redesign. They
are to cost 160,000 and 80,000 respectively and the work will get done for sure. The valve redesign
will be done before the software is developed. Expected development times have reduced and will
not be an issue for delivery.

The uncertainty now focuses only on the improvement that they can get from the New Valve. Rather
than the fuzzy estimate of modest/dramatic, a performance measure for evaluating valve-system
performance was going to be used. This scale rated valves between 200 and 800. The existing old
valve SV44A-10 was rated at 522. Harrington proposed that the scale be specified in the contract to
measure performance, and suggested that any improvement rating above 600 is dramatic.

Harrington believed that the estimated rating on the new valve would be anywhere between 525
and 700, and she thought that a rating around 625 was most likely. She also indicated that the
actual rating, which would depend on how well the valve works with the software would be normally
distributed around her estimated rating value, with a standard deviation of 20 points.

After a lengthy conversation with Andre Gide of Avion, the two agreed that, for every point above 522
and up to 600, the price of each valve system would increase by $25 a point over the $10,000 price
of the old valve. If the improvement rating exceeded 600, in addition to the extra $1,950 for the first
78 points of improvementArmstrong proposed $200 a point for each point above 600. Gide, after
some reflection, agreed to this arrangement, with the proviso that the total price per valve could not
exceed $25,000. Armstrong also recalled that it costs $8000 per valve to manufacture the old valve,
and that it would cost $10,500 to manufacture the new valve.

a. Assume that you will develop and sell the new valve to Avion no matter what estimated rating
Harrington provides. Develop a simple @Risk model to estimate the profit (for the sale of the 50
values) associated with this strategy

b. Report the expected value and the standard deviation associated with the development of this new
valve. What is the probability that the total profit being more than $100,000? Would you recommend
producing the new valve or the old valve (which has a guaranteed profit of $100,000)?

c. Suppose you consider a new strategy of developing a cut off value. The idea is that if
Harringtons estimated rating is less than this cut off, you abandon the new valve and just pay the
$80,000 in engineering costs but save the $160,000 in software costs, and you then manufacture the
old valve. Modify your simulation model above to incorporate this.
d. Report the expected value and the standard deviation associated with the development of this new
valve using this strategy. Try various cut-offs from 535-675 in increments of 20. What cut off values
would you recommend based on averages?

e. If Armstrong had the utility u(x)=1-e(-x/50000) then which cut-off would you recommend?